extensions: optionally print hint on import failure Test will be added by the next patch.

error: add hint to ProgrammingError As the hint isn't shown by the default exception handler, we need to print it manually. I've copied the "** " style from _exceptionwarning().

registrar: unindent superfluous "if True" block

registrar: switch @command decorator to class It overrides _funcregistrarbase._doregister() since the structure of the command table is quite different.

registrar: move cmdutil.command to registrar module (API) cmdutil.command wasn't a member of the registrar framework only for a historical reason. Let's make that happen. This patch keeps cmdutil.command as an alias for extension compatibility.

gendoc: make sure locale path is set before loading any modules Otherwise some messages wouldn't be translated depending on when the util was loaded.

fsmonitor: don't attempt state-leave if we didn't state-enter The state-enter command may not have been successful; for example, the watchman client session may have timed out if the user was busy/idle for a long period during a merge conflict resolution earlier in processing a rebase for a stack of diffs. It's cleaner (from the perspective of the watchman logs) to avoid issuing the state-leave command in these cases. Test Plan: ran `hg rebase --tool :merge -r '(draft() & date(-14)) - master::' -d master` and didn't observe any errors in the watchman logs or in the output from `watchman -p -j <<<'["subscribe", "/data/users/wez/fbsource", "wez", {"expression": ["name", ".hg/updatestate"]}]'`

fsmonitor: acquire localrepo.wlock prior to emitting hg.update state we see some weird things in the watchman logs where the mercurial process is seemingly confused about which hg.update state it is publishing through watchman. On closer examination, we're seeing conflicting pids for the clients involved and this implies a race. To resolve this, we extend the wlock around the state-enter/state-leave events that are emitted to watchman. Test Plan: Some manual testing: In one window, run this, and then checkout a different rev: ``` $ watchman -p -j <<<'["subscribe", "/data/users/wez/fbsource", "wez", {"expression": ["name", ".hg/updatestate"]}]' { "version": "4.9.0", "subscribe": "wez", "clock": "c:1495034090:814028:1:312576" } { "state-enter": "hg.update", "version": "4.9.0", "clock": "c:1495034090:814028:1:312596", "unilateral": true, "subscription": "wez", "metadata": { "status": "ok", "distance": 125, "rev": "a1275d79ffa6c58b53116c8ec401c275ca6c1e2a", "partial": false }, "root": "/data/users/wez/fbsource" } { "root": "/data/users/wez/fbsource", "metadata": { "status": "ok", "distance": 125, "rev": "a1275d79ffa6c58b53116c8ec401c275ca6c1e2a", "partial": false }, "subscription": "wez", "unilateral": true, "version": "4.9.0", "clock": "c:1495034090:814028:1:312627", "state-leave": "hg.update" } ``` Tailed the watchman log file and looked for invalid state assertion errors, then ran my `rebase-all` script to update/rebase all of my heads. Didn't trigger the error condition (but couldn't reliably trigger it previously anyway), and the output captured above shows that the states are being emitted correctly.

obsolete: move the 'isenabled' function at the top of the file That is a simple and important function so having it at the top next to the related constant seems better.

cache: make the cache updated callback easily accessible to extension This will help extension to benefit from this new logic. As a side effect this clarify the 'transaction' method a little bit.

obscheckhead: skip context creation to test phase On repository with many many heads, this is called often enough to show an overhead to appears in some profile. So we skip the 'changectx' creation to test the node phases.

tests: add missing 'test-push-checkheads-partial-C1.t' For some reason, this one test case did not made it with the others as they got upstreamed during the 4.2 cycle.

contrib: add editmerge version for powershell This just adds a translation of existing contrib/editmerge to powershell. It allows users on Windows to iteratively resolve conflicts in their editor of choice. # no-check-commit

run-tests: allow hg executable to be hg.exe When running tests on Windows (via msys), user sometimes does not want to run them against source hg, but against compiled hg.exe. For that purpose, --with-hg option can be used, but currently run-tests.py prints a warning if the value of this argument is not a file with basename 'hg'. This patch allows such file to be 'hg.exe'.

obsolete: add operation metadata to rebase/amend/histedit obsmarkers By recording what operation created the obsmarker, we can show very intuitive messages to the user in various UIs. For instance, log output could have messages like "Amended as XXX" to show why a commit is old and has an 'x' on it. @ ac28e3 durham / First commit | | o d4afe7 durham | | Second commit | | | x 8e9a5d (Amended as ac28e3) durham |/ First commit |

match: use match.prefix() in subdirmatcher It seems like the subdirmatcher should be checking if the matcher it's based on is matching prefixes. It was effectively doing that already because "prefix() == not always() and not anypats() and not isexact()", subdirmatcher was checking the first two parts of that condition and I don't think it will ever be given an "exact" matcher with it's directory name (because exact matchers are for matching files, not directories). Still, let's switch to using prefix() for clarity (and because I'm trying to remove code that reaches for matchers internals).

match: avoid accessing match._pathrestricted from subdirmatcher Accessing only the public API wherever possible helps us refactor matchers later.

match: override visitdir() the usual way in subdirmatcher Just override the function instead of replacing it on each instance.

match: make _fileroots a @propertycache and rename it to _fileset The files in the set are not necesserily roots of anything. Making it a @propertycache will help towards extracting a base class for matchers.

fsmonitor: use modern match API It seems like fsmonitor/__init__.py was based on a pretty old version of dirstate.py. Let's copy over the changes from the following two commits: 55c449345b10 (match: add isexact() method to hide internals, 2014-10-29) 3c346969c321 (dirstate: avoid match.files() in walk(), 2015-05-19)

diffstat: properly count lines starting in '--' or '++' (issue5479) Lines that start in '--' or '++' were previously not counted as deletions or additions in diffstat, resulting in incorrect addition/deletion counts. The bug was present if the start of the line, combined with the diff character resulted in '---' or '+++'. diffstatdata will now track, for each file, if it has moved pas the header section by looking for a line beginning with '@@'. Once that has happened, lines beginning with '-' or '+' will be counted for deletions and additions. Once a line beginning with 'diff' is found, the process starts over.

test-http-bad-server: match different in output Same as for previous update, the longer port number can affect this ouput.

config: make config.items() return a copy config.items() was iterating over a copy of the data for the the specified section on Python 2 by using .items(). However, on Python 3, items() does not make a copy, so let's switch to explicitly making a copy to make it safe on both Python 2 and Python 3.

filemerge: store error messages in module variables Copytracing may be disabled because it's too slow (see experimental.disablecopytrace config option). In that case user may get errors like 'local changed FILE which other deleted'. It would be nice to give user a hint to rerun command with `--config experimental.disablecopytrace=False`. To make it possible let's extract error message to variables so that extension may overwrite them.

run-tests: support multiple cases in .t test Sometimes we want to run similar tests with slightly different configurations. Previously we duplicate the test files. This patch introduces special "#testcases" syntax that allows a single .t file to contain multiple test cases. Defined cases could be tested using "#if". For example, if a test should behave the same with or without an experimental flag, we can add the following to the .t header: #testcases default experimental-a #if experimental-a $ cat >> $HGRCPATH << EOF > [experimental] > feature=a > EOF #endif The "experimental-a" block won't be executed when running the "default" test case.

revlog: rename constants (API) Feature flag constants don't need "NG" in the name because they will presumably apply to non-"NG" version revlogs. All feature flag constants should also share a similar naming convention to identify them as such. And, "RevlogNG" isn't a great internal name since it isn't obvious it maps to version 1 revlogs. Plus, "NG" (next generation) is only a good name as long as it is the latest version. Since we're talking about version 2, now is as good a time as any to move on from that naming.

localrepo: reformat set literals Putting multiple elements on the same line makes diffs harder to read. Switch to one line per element so future changes are easier on the eyes.

rebase: make sure merge state is cleaned up for no-op rebases (issue5494) If a rebase ends up doing a no-op commit, make sure the merge state is still cleaned up.

match: remove ispartial() The function was added in d1d69ca78883 (match: add match.ispartial(), 2015-05-15) for use by narrowhg, but narrowhg never ended up needing it.

run-tests: rename test description dict to testdesc Previously the word "test" was used for both a Test instance and a path or test dict. This patch renames them so it's clear that "testdesc" is the dict, and "test" is the instance.

run-tests: change test identity from a path to a dict Previously, we use path to identify a test. A later patch adds more information so a path is not enough to identify a test. So we change it to a dictionary.

largefiles: delete unnecessary meddling with matcher internals lfutil.getstandinmatcher() was setting match._always to False because it wanted a matcher of no patterns to match no files and match.match() instead matches everything. However, since ba8089433090 (largefiles: ensure lfutil.getstandinmatcher() only matches standins, 2015-08-12), it never actually passes an empty list of patterns, so the hack has become unnecessary.

py3: explicitly convert a list to bytes to pass in ui.debug Here pats is a list obviously. Since we can't pass unicodes to ui.debug, we have to pass this as bytes.

changelog: make sure datafile is 00changelog.d (API) 0ad0d26ff7 makes it possible for changelog datafile to be "00changelog.i.d", which is wrong. This patch adds an explicit datafile parameter to fix it.

util: drop unneeded override, sortdict.copy() OrderedDict.copy() returns self.__class__(self).

drawdag: provide filenode for its dummy filectx basefctx has filenode, which will be used in a future fast path of skipping flag processor during repo._filecommit. This patch adds filenode to drawdag's faked filectx so we don't need to use hasattr in localrepo.py.

tests: fix up recent conditionalized output changes It looks like (!) can have surprising results matching back to the original output when adjacent lines change, probably because it uses the same code matching that allows (?) to skip missing output. 24f55686a63d ended up adding unconditionalized check*{exec,link} lines, duplicating the conditionalized lines. A Windows run wanted to delete the unconditionalized lines. This now runs on both Windows and Linux.

run-tests: drop fallback for proc.terminate() for pre-py2.6

run-tests: drop fallback for sysconfig for pre-py2.7

largefiles: move identical statement to after if/else

util: rewrite sortdict using Python 2.7's OrderedDict Pattern copied from https://docs.python.org/dev/library/collections.html#ordereddict-examples-and-recipes.

encoding: use i.startswith() instead of i[0] to eliminate py2/3 divergence

test-casefolding: sync with latest code changes This goes with c2380b448265.

perf: always pass node to revlog.revision() I removed this in 73c3e226d2fc thinking it wasn't necessary. In fact, we need to always pass a node so the code is compatible with revisions before d7d64b89a65c. The new code uses a variable to avoid check-style complaining about "r.revision(r.node(" patterns.

hg: backout optimizing for treemanifests It turns out that the files list is not sufficient to identify with revlogs have changed. In a merge commit, no files could've changed but directories would have. For now let's just backout this optimization.

test-hardlink: do not test .hg/cache/checklink Linux and BSD have different behavior on "os.link(src, dst)" where "src" is a symlink. That causes test difference. According to POSIX [1]: If path1 names a symbolic link, it is implementation-defined whether link() follows the symbolic link, or creates a new link to the symbolic link itself. So both behaviors are correct. This patch removes the trouble maker "checklink" to make the test pass on both platforms. [1]: http://pubs.opengroup.org/onlinepubs/9699919799/functions/link.html

test-hardlinks: unify two test files into one The "whitelisted" test could be unified using "#if"s.

style: ban [ foo == bar] bashism in tests

changelog: load pending file directly When changelogs are written, a copy of the index (or inline revlog) may be written to an 00changelog.i.a file to facilitate hooks and other processes having access to the pending data before it is finalized. The way it works today, the localrepo class loads the changelog like normal. Then, if it detects a pending transaction, it asks the changelog class to load a pending changelog. The changelog class looks for a 00changelog.i.a file. If it exists, it is loaded and internal data structures on the new revlog class are copied to the original instance. The existing mechanism is inefficient because it loads 2 revlog files. The index, node map, and chunk cache for 00changelog.i are thrown away and replaced by those for 00changelog.i.a. The existing mechanism is also brittle because it is a layering violation to access the data structures being accessed. For example, the code copies the "chunk cache" because for inline revlogs this cache contains the raw revision chunks and allows the original changelog/revlog instance to access revision data for these pending revisions. This whole behavior of course relies on the revlog constructor reading the entirety of an inline revlog into memory and caching it. That's why it is brittle. (I discovered all this as part of modifying behavior of the chunk cache.) This patch streamlines the loading of a pending 00changelog.i.a revlog by doing it directly in the changelog constructor if told to do so. When this code path is active, we no longer load the 00changelog.i file at all. The only negative outcome I see from this change is if loading 00changelog.i was somehow facilitating a role. But I can't imagine what that would be because we throw away its data (the index data structures are replaced and inline revision data is replaced via the chunk cache) and since 00changelog.i.a is a copy of 00changelog.i, file content should be identical, so there should be no meaninful file integrity checking at play. I think this was all just sub-optimal code.

cleanup: use set literals We no longer support Python 2.6, so we can now use set literals.

py3: convert date and format arguments str before passing in time.strptime time.strptime() raises ValueError if the arguments are not str. Source Code: https://hg.python.org/cpython/file/3.5/Lib/_strptime.py#l307

py3: convert kwargs' keys to str using pycompat.strkwargs

verify: add a config option to skip certain flag processors Previously, "hg verify" verifies everything, which could be undesirable when there are expensive flag processor contents. This patch adds a "verify.skipflags" developer config. A flag processor will be skipped if (flag & verify.skipflags) == 0. In the LFS usecase, that means "hg verify --config verify.skipflags=8192" will not download all LFS blobs, which could be too large to be stored locally. Note: "renamed" is also skipped since its default implementation may call filelog.data() which will trigger the flag processor.

changegroup: add bundlecaps back Commit 282b288aa20c333c removed the unused bundlecaps argument from the changegroup code. While it is unused in core Mercurial, it was an important feature for the remotefilelog extension because it allowed the exchange layer to communicate to the changegroup packer that this was a shallow repo and that filelogs should not be included. Without bundlecaps, there is currently no other way to pass that information along without a more extensive refactor of exchange, bundle, and changegroup code. This patch backs out the original removal, and merges it with some recent changes to changegroup apis.

flagprocessor: add a fast path when flags is 0 When flags is 0, _processflags could be a no-op instead of iterating through the flag bits.

shelve: make shelvestate use simplekeyvaluefile Currently shelvestate uses line ordering to differentiate fields. This makes it hard for extensions to wrap shelve, since if two alternative versions of code add a new line, correct merging is going to be problematic. simplekeyvaluefile was introduced fot this purpose specifically. After this patch: - shelve will always write a simplekeyvaluefile - unshelve will check the first line of the file for a version, and if the version is 1, will read it in a position-based way, if the version is 2, will read it in a key-value way As discussed with Yuya previously, this will be able to handle old-style shelvedstate files, but old Mercurial versions will fail on the attempt to read shelvedstate file of version 2 with a self-explanatory message: 'abort: this version of shelve is incompatible with the version used in this repo'

shelve: refactor shelvestate loading This is a preparatory patch which separates file reading from the minimal validation we have (like turning version into int and checking that this version is supported). The purpose of this patch is to be able to read statefile form simplekeyvaluefile, which is implemented in the following patch.

extdiff: copy back execbit-only changes to the working directory Some tools like BeyondCompare allow the file mode to be changed. The change was previously applied if the content of the file changed (either according to size or mtime), but was not being copied back for a mode-only change. That would seem to indicate handling this in an 'elif' branch, but I opted not to in order to avoid copying back the mode without the content changes when mtime and size are unchanged. (Yes, that's a rare corner case, but all the more reason not to have a subtle difference in behavior.) The only way I can think to handle this undetected change is to set each file in the non-wdir() snapshot to readonly, and check for that attribute (as well as mtime) when deciding to copy back. That would avoid the overhead of copying the whole file when only the mode changed. But a chmod in a diff tool is likely rare. See also affd753ddaf1.

tests: remove regular expression matching for Python 2.6 This effectively reverts 52cca17ac523. Some lines still have (re) due to variable length port numbers. There's not much we can do about that. But at least this change removes most of the ugliness.

branchmap: remove use of buffer() to support Python 2.6 The use of buffer() was added in 7359157b9e46 to support Python 2.6, which we no longer support.